[HTCondor-users] Concurrency limit does not work on primary group

[Gang Qin <Gang.Qin@xxxxxxxxxxxxx>]

Dear Condor experts:

  Our negotiator is running condor-8.2.2-265643.x86_64, in our current configuration, we set  ConcurrencyLimits = strcat(AcctGroup, ",", AcctSubGroup, ",", Owner)

svr011:~# condor_q -global -autoformat ConcurrencyLimits | grep group_NONLHC  | tail -n 1
group_NONLHC,none,ilc023

'group_NONLHC' is the AcctGroup which is the primary group.
'none' is the AcctSubGroup which is a subgroup in 'group_NONLHC'
'ilc023' is a Owner in the subgroup 'none'.


Currently we have 453 ilc023 jobs running and 420 ilc jobs idle in the queue:
svr019:~# condor_ls --vo
Owner            Idle  Running  Held  Suspended Completed Removed Transferring_output
ilc023         420     453     4     0     0     0     0

 When I add a line 'ILC023_LIMIT = 400' and reconfigured the Negotiator,  in NegotiatorLog I can see:

     Rejected 183823.0 group_NONLHC.none.ilc023@xxxxxxxxxxxxxxx <10.141.255.10:60599>: concurrency limit none reached

 When I add a line 'NONE_LIMIT = 400' and reconfigured the Negotiator ,  in NegotiatorLog I can see:

     Rejected 186612.0 group_NONLHC.none.ilc023@xxxxxxxxxxxxxxx <10.141.255.10:60599>: concurrency limit none reached

  But when I add  a line 'GROUP_NONLHC = 400' and reconfigured the Negotiator ,  in NegotiatorLog I can only see:

     Rejected 79352.0 group_NONLHC.none.ilc023@xxxxxxxxxxxxxxx <10.141.255.11:46371>: no match found

  This message means the job has already passed the concurrency limit but not able to find free node.

  Any idea why the concurrency does not work on the primary group?

  Cheers,Gang